Raspberry pi based humanoid entity robot control system

Abstract

The application relates to a Raspberry Pi-based humanoid entity robot control system, which comprises a communication unit for receiving control instructions from an upper-layer application and parsing the control instructions into action instructions; a control unit connected with the communication unit, for receiving the action instructions, and performing semantic coding according to the action instructions to obtain an action semantic coding stream; an action unit connected with the control unit, for receiving the action semantic coding stream, decoding the received action semantic coding stream to obtain an action code stream, and calculating execution action data according to a control algorithm; and an execution unit connected with the action unit, the execution unit also being connected with motor and steering wheel control on the humanoid entity robot, and the execution unit being used for receiving the execution action data and controlling corresponding motors and steering wheels to execute the received execution action data. The application solves the problem that there is no Raspberry Pi-based humanoid entity robot control system at present, and fills the technical blank.

Description

Technical Field

[0001] This invention relates to the field of robotics, and specifically to a humanoid robot control system based on Raspberry Pi. Background Technology

[0002] Humanoid robots are robots designed and manufactured to mimic human form and behavior, typically possessing human-like limbs and a head, one or both. Robots are generally designed in different shapes to meet various application needs, such as robotic arms for industrial use, wheelchair robots, and walking robots. Humanoid robot research integrates multiple sciences, including mechanics, electronics, computer science, materials science, sensors, and control technology, representing a country's high-tech development level. Judging from the current state of robotics and artificial intelligence research, there is still a long way to go to fully realize highly intelligent and flexible humanoid robots. Furthermore, humanity's incomplete understanding of itself further limits the development of humanoid robots.

[0003] Humanoid robots have a human appearance, can adapt to human living and working environments, replace humans in performing various tasks, and can extend human capabilities in many ways. They are widely used in many fields such as service, medical care, education, and entertainment.

[0004] The task of the control system is to control the robot's actuators according to the robot's operational instructions and signals fed back from sensors, enabling it to complete the prescribed movements and functions. If the robot lacks feedback capabilities, the control system is called an open-loop control system; if the robot possesses feedback capabilities, the control system is called a closed-loop control system. This part mainly consists of computer hardware and control software. The software primarily comprises the human-machine interface system for communication between humans and the robot, as well as control algorithms. This part acts analogously to the human brain.

[0005] The Raspberry Pi, a microcomputer about the size of a credit card, was designed for computer programming education. Despite its small size, it possesses all the basic functions of a PC. Designing a robot control system based on the Raspberry Pi could make the robot correspondingly small, lightweight, and agile. Currently, there are no Raspberry Pi-based humanoid robot control systems on the market. Therefore, there is an urgent need to provide a Raspberry Pi-based humanoid robot control system. Summary of the Invention

[0006] The purpose of this invention is to overcome the shortcomings of the prior art and provide a humanoid robot control system based on Raspberry Pi, thereby solving the problem that there is currently no control system for humanoid robots based on Raspberry Pi.

[0007] The technical solution to achieve the above objectives is:

[0008] The application provides a Raspberry Pi-based humanoid entity robot control system, comprising:

[0009] A communication unit is configured to establish a communication link with an upper-layer application, receive control instructions from the upper-layer application, and parse the control instructions into action instructions.

[0010] A control unit connected to the communication unit is configured to receive the action instructions sent by the communication unit, and perform semantic coding according to the action instructions to obtain an action semantic coding stream.

[0011] An action unit connected to the control unit is configured to receive the action semantic coding stream sent by the control unit, decode the received action semantic coding stream to obtain an action code stream, and calculate execution action data according to a control algorithm.

[0012] The Raspberry Pi-based humanoid entity robot control system of the application solves the problem that there is no such control system at present, and fills the technical gap.

[0013] The further improvement of the Raspberry Pi-based humanoid entity robot control system of the application is that the action unit comprises a motor actuator module connected to the actuator.

[0014] The motor actuator module obtains the action code stream, calculates a cubic spline function with the obtained action code stream as sampling points, and calculates the interpolation results of each step point by bringing a linear coefficient compensation factor into the cubic spline function to obtain the execution action data sent to the actuator.

[0015] The further improvement of the Raspberry Pi-based humanoid entity robot control system of the application is that the action unit further comprises a servo actuator module connected to the actuator.

[0016] The servo actuator module obtains the action code stream, calculates a mean linear function with the obtained action code stream as sampling points, and obtains the interpolation results of each step point as the execution action data sent to the actuator.

[0017] The further improvement of the Raspberry Pi-based humanoid entity robot control system of the application is that the actuator is further configured to form real-time feedback data and feed back the real-time feedback data to the action unit.

[0018] The action unit is further configured to receive the real-time feedback data and feed back the real-time feedback data to the control unit.

[0019] The control unit is further configured to analyze the received real-time feedback data to obtain an analysis result, and send the analysis result to the communication unit.

[0020] The communication unit is further configured to send the analysis result to an upper application.

[0021] The further improvement of the human-like entity robot control system based on Raspberry Pi is that the action library is further included.

[0022] The control unit is connected with the action library, and the control unit queries the action library after receiving the action instruction, and performs semantic coding on the queried action to obtain an action semantic coding stream.

[0023] The further improvement of the human-like entity robot control system based on Raspberry Pi is that the communication unit includes a Bluetooth module, a data processing module and a serial port.

[0024] The Bluetooth module establishes a communication link with the upper application.

[0025] The data processing module is configured to parse the received control instruction to form an action instruction.

[0026] The communication unit is in communication connection with the control unit through the serial port.

[0027] The further improvement of the human-like entity robot control system based on Raspberry Pi is that the control unit sets a unified reference time axis when performing semantic coding.

[0028] The joint action is divided into multiple interval segment actions.

[0029] The actuator time of the longest time action point is set as the end time of the joint action, and the tn information of all actuators is filled according to the joint action.

[0030] When the end time of the joint action comes, the coding action is ended, and the action semantic coding stream is output.

[0031] The further improvement of the human-like entity robot control system based on Raspberry Pi is that the action unit sets a unified reference time axis when performing decoding.

[0032] The action semantic coding stream of all joint actions is read, the action code stream of the tn time point is taken out, the complete continuous smooth action point is calculated according to the action code stream by using an action smoothing algorithm, and the action code stream of the tn time point is executed.

[0033] When the end time of the joint action comes, the decoding action is ended. BRIEF DESCRIPTION OF DRAWINGS

[0034] Figure 1 System diagram of the present application based on Raspberry Pi humanoid entity robot control system.

[0035] Figure 2 System framework diagram of the present application based on Raspberry Pi humanoid entity robot control system.

[0036] Figure 3 Framework diagram of coding and decoding in the present application based on Raspberry Pi humanoid entity robot control system.

[0037] Figure 4 Curve diagram of the result obtained after obtaining the coefficients of cubic spline function in the present application.

[0038] Figure 5 Curve diagram of the speed obtained in the present application.

[0039] Figure 6 Curve diagram of the acceleration obtained in the present application. DETAILED DESCRIPTION

[0040] The present application will be further described below in combination with the drawings and specific embodiments.

[0041] Referring to Figure 1 , the present application provides a Raspberry Pi-based humanoid entity robot control system to solve the problem that there is no existing Raspberry Pi-based humanoid entity robot control system to fill the technical gap. The Raspberry Pi is small in size, and the control system designed by using the Raspberry Pi can make the robot also small, light and flexible in action. The present application based on Raspberry Pi humanoid entity robot control system will be described below in combination with the drawings.

[0042] Referring to Figure 1 , the system diagram of the present application based on Raspberry Pi humanoid entity robot control system is shown. The present application based on Raspberry Pi humanoid entity robot control system will be described below in combination with Figure 1 .

[0043] As Figure 1As shown, the Raspberry Pi-based humanoid entity robot control system of the present application comprises a communication unit 21, a control unit 22 and an action unit 23, the communication unit 21 is used to establish a communication link with an upper application and receive control instructions from the upper application, and parse the control instructions into action instructions; the control unit 22 is connected with the communication unit 21, and the control unit 22 is used to receive the action instructions sent by the communication unit 21, and perform semantic coding according to the action instructions to obtain an action semantic coding stream; the action unit 23 is connected with the control unit 22, and the action unit 23 is used to receive the action semantic coding stream sent by the control unit 22, and decode the received action semantic coding stream to obtain an action code stream, and calculate execution action data according to a control algorithm; the action unit 23 is also connected with an actuator of a humanoid entity robot for sending the execution action data to the actuator for execution.

[0044] In combination Figure 2 As shown, the upper application is an application control host computer or a mobile phone APP software, which can send control instructions to the humanoid robot control system; the communication unit runs in the system in the form of a communication process, processes the control instructions into action instructions after receiving the control instructions, and then sends the action instructions to the control unit; the control unit runs in the system in the form of a control process, queries an action semantic coding stream from a local action library after receiving the action instructions, and then outputs the action semantic coding stream to the action unit; the action unit runs in the system in the form of an action process, receives the action semantic coding stream and decodes the action semantic coding stream. Then, the execution action data is obtained by processing using a plurality of actuator linkage control algorithms and sent to the actuator; the actuator is arranged on the humanoid entity robot, and the actuator can be a motor or a servo, etc. The motor of the current humanoid robot is mainly used to control the actions of the arms, chest, waist and legs of the humanoid robot, and the servo is mainly used to control the actions of the head, wrist, palm and fingers of the humanoid robot.

[0045] In one specific embodiment of the present application, after the control system of the present application is started, initialization is first performed, including process initialization, communication interface initialization and actuator initialization, then processes are created, mainly communication processes, control processes and action processes are created, then the processes are started to begin to accept the input of instructions, if there is no input of instructions, the data feedback process is entered, if there is input of instructions, the instructions are received and processed, and output to the actuator for action output, then the data feedback process is performed, and the whole process is returned to the instruction input step to continue to circulate.

[0046] In one specific embodiment of the present application, the action unit 23 comprises a motor actuator module, and the motor actuator module is connected with the actuator;

[0047] The motor actuator module acquires the action code stream, calculates a cubic spline function with the acquired action code stream as sampling points, and brings a linear coefficient compensation factor into the cubic spline function to calculate interpolation results of each step point as execution action data and sends the execution action data to the actuator.

[0048] Specifically, the motor actuator module first defines a minimum action time slice, then acquires the action code stream, judges whether the action time is over, and if the action is over, stops the action; if the action is not over, calculates coefficients a, b, c and d according to a cubic spline function f(x) = a*x^3 + b*x^2 + c*x + d with the input action code stream as sampling points, then substitutes the calculated coefficients into the cubic spline function to calculate interpolation results of each step point, executes the action of each motor according to the calculation results, and then returns to continue judging whether the action time is over, and the cycle is executed until the action time is over.

[0049] The linear coefficient compensation factor is used to obtain relatively accurate smooth interpolation points. The linear coefficient compensation factor is calculated by the following method: there is a linear relationship between the theoretically calculated spline interpolation results and the actual values, i.e. B(t) = m*V(t), the linear coefficient compensation factor m is calculated by a plurality of actual measurement values, M1 = B(t1) / V(t1), M2 = B(t2) / V(t2)…Mn = B(tn) / V(tn), m = (M1+M2+…+Mn) / n, and the linear coefficient compensation factor is added to the cubic spline function to obtain relatively accurate smooth interpolation points.

[0050] Principle of cubic spline interpolation algorithm:

[0051] Suppose there are the following nodes:

[0052] x: a = x0 < x1 < … < xn = b

[0053] y: y0 y1 … yn

[0054] The spline curve S(x) is a piecewise defined formula, given n+1 data points, there are n intervals, and the cubic spline equation satisfies the following conditions:

[0055] In each sub-interval [xi, xi+1] (i = 0, 1, …, n-1, x is increasing), S(x) = Si(x) is a cubic polynomial;

[0056] S(xi) = yi (i = 0, 1, …, n) is satisfied;

[0057] S(x), derivative S`(x), and second derivative S``(x) are continuous in the interval [a, b], i.e. the S(x) curve is smooth;

[0058] So n cubic polynomial segments can be written as:

[0059] S i (x) = a i +b i (x-x i )+c i (x-x i ) 2 +d i (x-xi) 3 , i = 0, 1, …, n-1

[0060] Where ai, bi, ci, di represent 4n unknown coefficients.

[0061] Cubic spline interpolation algorithm to solve:

[0062] Given:

[0063] a. n+1 data points [xi, yi], i = 0, 1, …, n

[0064] b. Each segment is a cubic polynomial function curve

[0065] c. Nodes to second order continuous

[0066] d. Left and right end point characteristics (natural boundary, fixed boundary, non-node boundary)

[0067] According to the fixed point, the coefficients in each spline curve equation are obtained, that is, the specific expression of each curve is obtained.

[0068] Interpolation and continuity:

[0069] S i (x i ) = y i

[0070] S i (x i+2 )y i+1 , where i = 0, 1, …, n-1

[0071] Differential continuity:

[0072] S′ i (x i+1 ) = S′ i+1 (x i+1 )

[0073] S″ i (x i+1 ) S″ i+1 (x i+1 ), where i = 0, 1, …, n-2

[0074] The differential of the spline curve:

[0075] S i (x) = a i +b i (x-x i )+c i (x-x i ) 2 +d i (x-x i ) 3

[0076] S′ i (x) = b i +2c i (x-x i )+3d i (x-x i ) 2

[0077] S″ i (x) = 2c i +6d i (x-x i )

[0078] Substitute the step size h i =x i+1 -x i into the conditions of the spline curve:

[0079] a. From S i (x i ) = y i (i = 0, 1,..., n-1), we obtain a i =y i

[0080] b. From S i (x i+1 ) = y i+1 (i = 0, 1,..., n-1), we obtain

[0081]

[0082] c. From S′ i (x i+1 ) = S′ i+1 (x i+1 ) (i = 0, 1,..., n-2), we obtain

[0083] S′ i (x i+1 ) = b i +2c i (x i+1 -x i )+3d i(x i+1 -x i ) 2 = b i + 2c i h + 3d i h 2

[0084] S′ i+1 (x i+1 ) = b i+1 + 2c i (x i+1 -x i+1 ) + 3d i (x i+1 -x i+1 ) 2 = b i+1

[0085] Thus we have:

[0086]

[0087] d. From S″ i (x i+1 ) = S″ i+1 (x i+1 ) (i = 0, 1,..., n-2) we have

[0088] 2c i + 6h i d i - 2c i+1 = 0

[0089] Let m i = S″ i (x i ) = 2c i , then

[0090] a. 2c i + 6h i d i - 2c i+1 = 0 can be written as: m i + 6h i d i - m i+1 = 0, we have

[0091]

[0092] b. Substitute ci, di into we have:

[0093]

[0094] c. Substitute bi, ci, di into Available:

[0095]

[0096] The free boundary end of the simulation robot does not receive any force to bend them, that is, S'' = 0. Specifically, m0 = 0 and m n = 0, so the free boundary is solved, and the equation group to be solved can be written as:

[0097]

[0098] After the coefficients of the cubic spline function are finally obtained, the results are:

[0099] P(t) = S(t) = a * t^3 + b * t^2 + c * t + d

[0100] The obtained curve is shown in Figure 4 .

[0101] Derivation of the position P(t) can obtain the velocity V(t) = P`(t) = 3a * t^2 + 2bt + c curve, for example, similar curves are shown in Figure 5 .

[0102] Derivation of the position V(t) can obtain the acceleration A(t) = P``(t) = 6a * t + 2b curve, for example, similar curves are shown in Figure 6 .

[0103] In one specific embodiment of the application, the action unit 23 further comprises a steering engine module, and the steering engine module is connected with the actuator;

[0104] The steering engine module obtains the action code stream, calculates the mean linear function with the obtained action code stream as the sampling point, obtains the interpolation result of each step point, and sends the execution action data to the actuator.

[0105] Specifically, the steering engine module first defines the minimum action time slice, then obtains the action code stream, judges whether the action time is over, if the action is over, it stops, if the action is not over, it calculates the coefficient v according to the mean linear function s = v * t with the input action code stream as the sampling point, calculates the interpolation result of each step point according to the calculated coefficient, executes the action of the steering engine according to the calculation result, and then returns to continue to judge whether the action time is over, and the cycle is executed until the action time is over.

[0106] Since the steering engine is usually designated to the target angle, the steering engine is rotated to the target angle at the maximum speed, so that the robot cannot show the soft and smooth action in the actual motion control process of the humanoid robot, and the smooth motion control algorithm of the steering engine is needed to realize the soft and smooth action of the robot, the steering engine motion smooth control algorithm: S(t) = a * t + b, the initial time of the given action is t0, the initial position is S0, the given termination time is tn and the termination position is Sn,

[0107] Then a = (Sn-S0) / (tn-t0)

[0108] S(t) = a * (t-t0) + S0

[0109] b = S0-a*t0

[0110] Finally, the corresponding position interpolation is obtained according to the function calculated above and each minimum time slice point, so that the relatively smooth steering engine motion trajectory can be obtained.

[0111] In one embodiment of the present application, the actuator is further configured to feed the real-time feedback data to the action unit;

[0112] The action unit is further configured to receive the real-time feedback data fed to the control unit;

[0113] The control unit is further configured to analyze the received real-time feedback data to obtain an analysis result, and send the analysis result to the communication unit;

[0114] The communication unit is further configured to send the analysis result to the upper application.

[0115] As shown in Figure 2 The actuator feeds the current state information to the action process in real time, the action process feeds the real-time data to the control process, the data acquisition and analysis module in the control process analyzes the real-time data feedback, and finally outputs the result to the communication process, and the communication process feeds the result data to the upper application.

[0116] In one embodiment of the present application, the communication unit includes a Bluetooth module, a data processing module and a serial port;

[0117] The Bluetooth module establishes a communication link with the upper application;

[0118] The data processing module is configured to parse the received control instruction to form an action instruction;

[0119] The communication unit is in communication connection with the control unit through the serial port.

[0120] Specifically, the communication process in the system starts to wait for instruction input after starting, if no feedback result is received, the received data is processed and fed back to the application layer, and then the system continues to wait for instruction input; if instruction input is received, the input instruction is parsed into an action instruction after processing, and then output to the control process through the serial port, if no feedback result is received, the system continues to wait for instruction input, if feedback result is received, the received data is processed and fed back to the application layer.

[0121] In one specific embodiment of the application, an action library is further included.

[0122] The control unit is connected with the action library, and after receiving the action instruction, the control unit queries the action library, and performs semantic coding on the queried action to obtain an action semantic coding stream.

[0123] Specifically, the control process in the system starts to wait for action instruction input after starting, if no action instruction is received, it is judged whether real-time feedback data is received, if no feedback result is received, the system continues to wait for instruction input, if feedback result is received, the received data is collected and analyzed and processed and fed back to the communication process, and then the system continues to wait for instruction input; if action instruction is received, the action library is queried, the queried action is semantically coded, and the action semantic coding is output, then if no feedback result is received, the system continues to wait for instruction input, if feedback result is received, the received data is collected and analyzed and processed and fed back to the communication process, and then the system continues to wait for instruction input.

[0124] Further, the action process in the system starts to wait for action semantic coding input after starting, if no action instruction is received, it is judged whether real-time feedback data is received, if no feedback result is received, the system continues to wait for instruction input, if feedback result is received, the received data is processed and fed back to the control process Task, and then the system continues to wait for instruction input; if action coding semantic coding is received, the received action coding stream is decoded, and then a plurality of executor linkage control algorithms are used to output the humanoid robot action data, and the data is transmitted to a single executor control algorithm for processing, then if no real-time feedback data is received, the system continues to wait for instruction input, if feedback result is received, the received data is processed and fed back to the control process Task, and then the system continues to wait for action semantic coding input.

[0125] As Figure 3As shown, the coding and decoding flow of the application is as follows: first, a unified time axis is established, and a sampling element action data structure and a coding element action data structure are set, so as to have a unified data structure standard and time standard when sampling and action coding, decoding and executing action code flow, then data sampling is performed on each actuator (motor or steering wheel, etc.) according to the sampling element action data structure, then the sampling points are coded according to the action semantics, the coding is completed, the coding is output and stored in the action library, when the robot moves, the action code flow needs to be taken out from the action library first, and the action code flow is decoded, then the continuous action code flow is calculated by using the spline interpolation algorithm, then the continuous action code flow is filled at each point according to the decoding element action data structure, and finally the decoding flow is output and the action code flow is executed according to the unified time axis.

[0126] Specifically, the sampling and coding element action data structure is S=[id, t, r], wherein S represents the sampling and coding data structure, id represents the actuator number, t represents the time point of the actuator running to the current angle, and r represents the current angle of the actuator. The decoding element action data structure A=[id, t, r, v, a], wherein A represents the decoding element action, id represents the actuator code, t represents the time point of the actuator running to the current angle, r represents the current angle of the actuator, v represents the current speed of the actuator, and a represents the current acceleration of the actuator. The joint element action data structure is B=[t, An], wherein B represents the joint action, t represents the unified time axis of the joint action, and An represents the current information of the N actuators.

[0127] Under the unified time axis t, the actuators numbered 1 to N rotate according to their respective action coding flows, the execution flows of the actuators are independent of each other and do not interfere with each other, so that the joints of the humanoid robot can move independently at the same time point, and finally the overall effect can realize the linkage action of the humanoid robot.

[0128] Further, when the control unit performs semantic coding, a unified reference time axis is set, and all actuator actions are based on the reference time axis;

[0129] The joint action is divided into multiple interval segment actions;

[0130] The actuator time of the longest time action point is set as the end time of the joint action, and each actuator fills the data of each interval segment action at the tn time point according to the element action A, and fills the tn information of all actuators according to the joint action B;

[0131] When the end time of the joint action comes, the coding action is ended, and the action semantic coding flow is output.

[0132] Further, the action unit sets a uniform reference time axis when decoding, and all executor actions are based on the reference time axis;

[0133] Read the action semantic code stream of all joint actions, take out the action code stream of the tn time point, calculate the complete continuous smooth action point according to the action code stream by using the action smoothing algorithm, and execute the action code stream of the tn time point; when the end time of the joint action comes, the decoding action is ended.

[0134] The above embodiments of the present application are described in detail with reference to the accompanying drawings, and those skilled in the art can make various changes to the present application according to the above description. Therefore, some details in the embodiments should not constitute a limitation on the present application, and the scope of protection of the present application will be defined by the appended claims.

Claims

1. A Raspberry Pi based humanoid entity robot control system, characterized in that, include: The communication unit is used to establish a communication link with the upper-layer application and receive control commands from the upper-layer application, and parse the control commands into action commands; A control unit connected to the communication unit is configured to receive action commands sent by the communication unit and perform semantic encoding based on the action commands to obtain an action semantic encoded stream; and The action unit connected to the control unit is used to receive the action semantic encoding stream sent by the control unit, decode the received action semantic encoding stream to obtain the action code stream, and then calculate the execution action data according to the control algorithm; the action unit is also connected to the actuator control of the humanoid robot, and is used to send the execution action data to the actuator for execution; The action unit includes a motor actuator module, which is connected to the actuator; The motor actuator module acquires the action code stream, uses the acquired action code stream as sampling points to calculate the cubic spline function, and substitutes the linear coefficient compensation factor into the cubic spline function to calculate the interpolation result of each step point as the execution action data and sends it to the actuator. The motor actuator module first defines the minimum action time slice, then acquires the action code stream, and determines whether the action time has ended. If the action has ended, the action stops; if the action has not ended, the input action code stream is used as sampling points to calculate coefficients a, b, c, and d according to the cubic spline function f(x) = a*x^3 + b*x^2 + c*x +d. Then, the calculated coefficients are substituted into the cubic spline function to calculate the interpolation result of each step point. Based on the calculation result, the action of each motor is executed, and then the module returns to continue to determine whether the action time has ended. This process is repeated until the action time ends. The linear coefficient compensation factor is calculated as follows: the theoretically calculated spline interpolation result has a linear relationship with the actual value, that is, B(t)=m*V(t), and the linear coefficient compensation factor m is calculated through multiple sets of actual measured values.

2. The Raspberry Pi based humanoid entity robot control system as claimed in claim 1 wherein, The actuation unit also includes a servo actuator module, which is connected to the actuator; The servo actuator module acquires the action code stream, uses the acquired action code stream as the sampling point to calculate the mean linear function, and obtains the interpolation result of each step point as the execution action data and sends it to the actuator.

3. The Raspberry Pi based humanoid entity robot control system as claimed in claim 1 wherein, The actuator is also used to generate real-time feedback data and feed it back to the action unit; The action unit is also used to feed back the received real-time feedback data to the control unit; The control unit is also used to analyze the received real-time feedback data to obtain analysis results, and send the analysis results to the communication unit; The communication unit is also used to send the analysis results to the upper-layer application.

4. The Raspberry Pi based humanoid entity robot control system of claim 1, wherein, It also includes an action library; The control unit is connected to the action library. After receiving an action instruction, the control unit queries the action library and performs semantic encoding on the queried actions to obtain an action semantic encoding stream.

5. The Raspberry Pi based humanoid entity robot control system as claimed in claim 1 wherein, The communication unit includes a Bluetooth module, a data processing module, and a serial port; The Bluetooth module establishes a communication link with the upper-layer application; The data processing module is used for parsing the received control instruction to form an action instruction; The communication unit is in communication connection with the control unit through a serial port.

6. The Raspberry Pi based humanoid entity robot control system of claim 1, wherein, The control unit sets a uniform reference time axis when performing semantic coding; The joint action is divided into multiple interval segment actions; The actuator time of the longest time action point is set as the end time of the joint action, and each actuator is filled with data according to the meta-action at the tn time point, and all actuator tn information is filled according to the joint action; When the end time of the joint action arrives, the coding action is ended, and an action semantic coding stream is output.

7. The Raspberry Pi based humanoid entity robot control system of claim 1, wherein, The action unit sets a uniform reference time axis when performing decoding; The action semantic coding stream of all joint actions is read, the action code stream at the tn time point is taken out, complete continuous smooth action points are calculated according to the action code stream by using an action smoothing algorithm, and the action code stream at the tn time point is executed; When the end time of the joint action arrives, the decoding action is ended.

Images